An unmet clinical need in breast cancer management is the accurate identification of patients who will benefit from adjuvant radiotherapy. We hypothesized that integration of postradiation clonogenic survival data with gene expression data across breast cancer cell (BCC) lines would generate a radiation sensitivity signature (RSS) and identify patients with tumors refractive to conventional therapy. Using clonogenic survival assays, we identified the surviving fraction (SF-2Gy) after radiation across a range of BCC lines. Intrinsic radiosensitivity was correlated to gene expression using Spearman correlation. Functional analysis was performed in vitro, and enriched biologic concepts were identified. The RSS was generated using a Random Forest model and was refined, cross-validated, and independently validated in additional breast cancer datasets. Clonogenic survival identifies a range of radiosensitivity in human BCC lines (SF-2Gy 77%-17%) with no significant correlation to the intrinsic breast cancer subtypes. One hundred forty-seven genes were correlated with radiosensitivity. Functional analysis of RSS genes identifies previously unreported radioresistance-associated genes. RSS was trained, cross-validated, and further refined to 51 genes that were enriched for concepts involving cell-cycle arrest and DNA damage response. RSS was validated in an independent dataset and was the most significant factor in predicting local recurrence on multivariate analysis, outperfoming all clinically used clinicopathologic features. We derive a human breast cancer-specific RSS with biologic relevance and validate this signature for prediction of locoregional recurrence. By identifying patients with tumors refractory to standard radiation this signature has the potential to allow for personalization of radiotherapy.